AI Ethics Disclosure

1. Transparency in AI Functionality: "The Glass Box" Approach

Unlike "black-box" corporate AI, Connected Classroom operates with ‘glass-box’ radical transparency approach. Our Intelligence Suite (REAL Connections, CrossLink, and Unautomatable AI) uses generative AI to augment, not automate, the instructional design process.

  • Pedagogical Logic:

    • REAL Connections: Synthesizes standards-aligned, authentic, and place-based applications using Authentic Learning Theory.

    • CrossLink: Utilizes a proprietary 5-Dimensional Scaffolding logic to identify rigorous, interdisciplinary intersections.

    • Unautomatable AI: Generates micro-lessons focused on uniquely human capacities (Empathy, Ethics, Critical Inquiry) that AI cannot replicate.

  • Model Specification: Our tools are powered by the Google Gemini API (gemini-flash-latest). We utilize custom system-level prompt engineering and educational fine-tuning via a secure serverless proxy to ensure outputs align with research-backed instructional strategies.

  • Human Oversight: We strictly adhere to a "Non-Automated Decision" policy. Every output is a pedagogical suggestion requiring the professional judgment and contextual expertise of a human educator.

2. Bias Mitigation: Research-Informed Guardrails

Corporate EdTech often "filters" bias after the fact. We mitigate bias at the architectural level through:

  • Scaffolding Logic: CrossLink’s logic-gate ensures that cognitive complexity and support density are developmentally aligned (e.g., distinguishing the needs of K-2 vs. 9-12), preventing "one-size-fits-all" algorithmic bias.

  • Manuscript Injections: Unautomatable AI uses "manuscript logic" to force the model to prioritize universally accepted human values and diverse perspectives in every output.

  • Equitable Representation: Our prompt engineering explicitly mandates balanced cultural, gender, and socioeconomic representation in case studies, historical figures, and community partners.

  • Feedback Loop: We maintain an open channel for "Algorithm Auditing." Educators can report perceived bias to ethics@connectedclassroom.org for manual logic adjustment.

3. Human-in-the-Loop (HITL) Design

We believe AI should be a "Thinking Catalyst," never the "Thinking Partner." Our design philosophy focuses on Educator Agency:

  • Suggestive, Not Prescriptive: All outputs are presented as modular drafts. The "Classroom-First" principle ensures that the educator, not the algorithm, is the final arbiter of what enters the classroom.

  • Modular Customization: Tools allow for "Deep Contextualization," enabling teachers to modify grade level, local priorities, and student interests, ensuring the AI serves the local community rather than a generic data set.

  • Professional Autonomy: Our training resources emphasize Collaborative Intelligence, teaching educators how to interrogate AI outputs rather than defer to them.

4. Privacy by Design: The Ephemeral Architecture

While corporate tools often monetize user data or use it to train their models, Connected Classroom operates on a Zero-Data Retention model.

  • The Zero-Data Principle: We do not collect, store, or profile teacher or student data. All sessions are ephemeral. This means your inputs are processed in real-time and purged immediately after the output is generated.

  • Secure Serverless Proxy: We utilize a Google Apps Script "Serverless Proxy" architecture. This means:

    • API Security: Our proprietary logic and API keys are never exposed to the client-side browser.

    • No Tracking: We do not use persistent cookies or trackers. We only log anonymous metadata (e.g., Tool Type, Grade Level) to a private audit sheet for system optimization.

    • Commercial Integrity: Your instructional ideas are your intellectual property. We will never sell, lease, or use your inputs for commercial gain or model training.

5. Accountability & Redress

We hold ourselves to the highest standards of professional EdTech ethics:

  • Rapid Corrective Action: We commit to investigating any report of inappropriate or inaccurate AI suggestions within 72 hours, followed by immediate updates to the underlying system instructions.

  • Open Governance: Our practices will be informed by an informal Ethics Advisory Board consisting of practicing K-12 educators and instructional designers.

  • Standards Alignment: Our ethics framework is built to exceed the requirements set by the ISTE Standards and the U.S. Department of Education (2023) recommendations on "AI and the Future of Teaching and Learning."